1. Simulation & Reconstruction Durga Rajaram IIT, Chicago MICE Project Board 29 April 2014

2. MPB-6 Recommendation Present a status report on the MICE simulation, online and offline analysis capabilities and show results of the end-to-end Monte Carlo simulations, including tracking and reconstruction, in support of the Step IV, V and VI physics goals at the next meeting. 29/4/14, MPB2

3. MICE Detectors 29/4/14, MPB3 TOF Tracker Ckov EMR KL

4. Organization Changes: EMR –Student taking over from Asfandiyarov, some associated learning curve Ckov –New Undergraduate students supervised by DR Global –Jan Greis replaced Lane –Ian Taylor leaving. Need experienced hand to guide. –Critical item 29/4/14, MPB4

5. Offline Flow 29/4/14, MPB5

6. Monte Carlo Beam Geometry Particle tracking Detector response 29/4/14, MPB6

7. Beam generation Previously: –Generate pencil beam, or sample from a user-defined gaussian, or read in from a file Now added G4Beamline as a generator –Updated survey, geometries and currents –Generates beam upto D2 –More realistic beam, better agreement with data (Step I) cf. Analysis presentation –Ability to generate beam based on data-run currents 29/4/14, MPB7 Nugent SimulationData

8. Geometry Default geometry description has been through flat text files –no surveying or versioning information –Prone to errors and outdatedness It was decided to have a CAD-based geometry description that is stored in the database –Significant progress –CAD models converted to GDML descriptions and stored in CDB –Realistic Step I & Idealistic Step IV geometries now available Except Ckov and EMR which are being finalized and integrated –Requires validation and shakedown –Geometry loading time is an issue especially for user-testing, tuning, studies, etc –This is a new system, so there is risk associated with its deployment and usability –“Legacy” geometry continues to be supported and will be until the new geometry is proven 29/4/14, MPB8 Bayes/Ricciardi

9. Simulation Particle tracking, energy loss, and scattering are done through GEANT –Upgraded to GEANT 4.9.6 – better handling of tessellated solids in geometry Custom field map models, or read in maps from file Two steps to simulating detector response 1.Collect hits in each sensitive volume – volume ID, energy deposit, hit position, momentum, time… 2.Electronics response aka digitization – mock DAQ readout – volume ID to cable map, energy to ADC… 29/4/14, MPB9

10. Detector Response (TOF) Stable –Energy deposited is first converted to photoelectrons and then to an ADC count –Time of the hit is propagated to PMTs and converted to a TDC count –Calibration corrections are added in so that they can be taken out at reconstruction stage as is done with data Potential restructuring for trigger simulation (TOF is experiment’s trigger) 29/4/14, MPB10 Rajaram

11. Detector Response (KL) New: Digitizer added along with updated geometry, and software to store hits in sensitive volumes –Outputs are ADC counts converted from energy deposits –Conversion factors are preliminary and likely require tuning –Needs validation against Step I data –Step I PID analysis can now use digitized and reconstructed KL 29/4/14, MPB11 Bogomilov

12. Detector Response (Tracker) Stable –Energy deposited is converted to photoelectrons, photoelectrons from a fiber channel are then summed and converted to an ADC New –Added Poisson noise to simulate VLPC dark current 29/4/14, MPB12 Dobbs/Heidt/Hunt/Santos

13. Detector Response (in progress) The Ckov and EMR digitizations are not yet in the production release of MAUS Ckov: –digitization depends on finalizing geometry (Cremaldi and student), simulating optical photons and collecting hits and digitizing (Rajaram and students). –Possible refactoring, or rewrite –Risk: undergraduate students who are new to this and have to balance effort with coursework EMR: –The final geometry and hit collection are now complete and in MAUS –Digitization is getting started 29/4/14, MPB13 CosmicMC

14. Reconstruction For any given detector, the reconstruction algorithm is required to be agnostic about input – should not distinguish data from MC At the moment, MAUS can reconstruct every single Step IV detector –TOF, Ckov, KL, Tracker, EMR –Some are mature, some functional with optimizations and tuning in progress, some preliminary The final Global reconstruction (in progress) will take the individual detector reconstructions and provide a global track and an associated particle identification hypothesis 29/4/14, MPB14

15. Reconstruction (TOF) Stable –Digits (MC or DAQ) are converted to slab hits. –x-y slab combinations are then used to form space- points –Space-point times are corrected for time-walk and trigger offsets based on calibrations stored in the DB RFI: Calibration algorithm needs some optimization to cover slabs/pixels with lower statistics –Reconstruction discards uncalibrated pixels resulting in a loss in acceptance 29/4/14, MPB15 Rajaram

16. Reconstruction (Ckov) Functional –Flash ADC samples integrated and converted to number of photoelectrons –Conversion factor requires tuning based on single-photoelectron fits and should be stored in DB (hard-coded now) –Charge integration window may need to be optimized –Add multi-peak reconstruction 29/4/14, MPB16 Drews/Winter/Rajaram π μ e

17. Reconstruction (Tracker) Functional –Space-point reconstruction, Pattern Recognition, Kalman tracking complete –Optimization underway –Efficiency studies in progress using MC –Calibration interface needs to be added Much improved documentation 29/4/14, MPB17 Dobbs/Heidt/Hunt/Santos

18. Reconstruction (EMR) New: Reconstruction of plane hits complete Higher level reconstruction (tracks, charge ratios, range) are being developed – to be integrated Asfandiyarov is moving on, and Drielsma is taking over, so there is some learning curve –Risk if Drielsma has to move on to graduate/do other things 29/4/14, MPB18 Asfandiyarov, Drielsma, Karadzhov

19. Reconstruction (Global) Global track fitting and particle ID are critical items –Tracking suffered from lack of experienced hand and grad student working on it has moved on. –A grad student (Jan Greis) has taken over and is getting up to speed –PID: framework developed & tested for TOF –In development: PID using TOF (time) + Tracker (momentum) –PID needs integration with the global tracking as that develops Ian Taylor, head of global reconstruction, is leaving  need someone experienced to guide the effort 9 months should be enough time, but requires fine breakdown to allow monitoring and catching risks early 29/4/14, MPB19 Greis/Pidcott/Taylor

20. Online Reconstruction Reconstruction software is same as for offline Distributed processing over worker nodes Summary and monitoring histograms visualized in MLCR as images via web interface –Plots intended to provide summary view of reconstruction quality –Lower level reconstruction plots serve to identify problems with detectors – e.g. inefficient tubes, dead channels… –Changes being made to visualization infrastructure (ROOT objects instead of raw images) Online reconstruction is available for TOF, Ckov, KL –Preliminary EMR display is available stand-alone, to be integrated with MAUS 29/4/14, MPB20

21. Software QA & Testing Software QA process –Style tests: does the code conform to style guidelines? –Unit tests: does the function do what’s expected? Currently, python coverage is ~68%, C++ coverage is ~78% –Integration tests: testing workflow – do all the pieces play nice together –Validation tests: does the simulation/reconstruction produce expected results? Continuous Integration test servers provide offline and online environments for developers to test their code before merging with the development branch Redmine issue tracker for bugs and features New release every ~2-4 weeks 29/4/14, MPB21

22. 29/4/14, MPB22 Schedule

23. 29/4/14, MPB23 Schedule

24. 29/4/14, MPB24 Schedule

25. 29/4/14, MPB25 Schedule

26. Conclusions MAUS can reconstruct data from every Step IV detector Capable of simulating the TOF, KL, and Tracker –EMR and Ckov in development, to be integrated The big gap is global reconstruction –Need new manager, and frequent monitoring to catch risks early CAD-geometry implementation has progressed –Needs validation against full simulation –Usability and optimization issues to be evaluated Need to focus and make sure we have working simulation and reconstruction (beam + geometry + detectors + global) ready for Step IV 29/4/14, MPB26